Well, the binomial distribution does not really fit here, as it measures the number of successes without placing any weight on losses.
This is wrong. It fits perfectly as there is only two possible outcomes win or loss and when n is large a reasonable approximation to it is given by the normal distribution.